Do local landscape features affect wild pollinator abundance, diversity and community composition on Canterbury farms? - New Zealand Ecological ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
262 New Zealand Journal of Ecology, Vol. 42, No. 2, 2018 DOI: 10.20417/nzjecol.42.29 Do local landscape features affect wild pollinator abundance, diversity and community composition on Canterbury farms? Kristina J. Macdonald*, Dave Kelly and Jason M. Tylianakis Centre for Integrative Ecology, School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand. *Author for correspondence (Email: Kristina.Macdonald@ccc.govt.nz) Published online: 20 June 2018 Abstract: Pollination is an essential ecosystem service that can be affected by habitat features in the immediate environment, termed here ‘local landscape features’. This study tested how five local landscape features (bare ground, native biodiversity plantings, homestead gardens, shelterbelts, and control areas of pasture) affect local pollinator communities on Canterbury farms. We also compared two sampling methods (flower visitation to native potted plants vs sticky traps) to determine if the sampling method affects the results of landscape-feature comparisons. We recorded 928 pollinators of 17 taxa on the potted plants and 791 pollinators of 16 taxa on the sticky traps. There were significant differences in pollinator abundances between the landscape features and the control areas. Both sampling methods recorded fewer pollinators overall at shelterbelts and bare ground sites than control sites, although morning-evening fly was more common at bare ground sites. However, the methods gave contrasting results for biodiversity plantings and gardens: the flower method recorded significantly more pollinators in biodiversity plantings and gardens than it did at control sites, whereas the sticky trap method recorded significantly fewer. As the flower method showed higher pollinator abundances near biodiversity plantings and gardens, planting native insect-pollinated plants on farms could boost populations of wild pollinators, which may improve crop pollination. Keywords: agriculture; bees; biodiversity; habitat; insects; pollination Introduction Therefore, local management decisions could be vital for wild pollinators and, in turn, insect-pollinated cropping systems. Pollination is essential for the reproduction and survival Habitat features in local environments can have varying of many wild plant populations, cultivated gardens and impacts on wild pollinator communities. Garibaldi et al. (2014) agricultural systems (De Groot et al. 2002). In fact, 75% suggest that key habitat for wild pollinators should include of global crops depend on insect pollination (Klein et al. semi-natural areas, habitat heterogeneity, nesting resources 2007). The intensification of food crop production is likely and hedgerows. The proximity to and area of semi-natural and increasing the requirements for pollination services per land natural vegetation has been associated with higher abundances area and simultaneously reducing local and regional pollinator and diversity of bee species (Kremen et al. 2002; Klein et al. biodiversity (Ghazoul, 2005). 2003; Ricketts et al. 2008; Garibaldi et al. 2011; Blaauw & There are global concerns about pollinator decline Isaacs 2014). Agricultural intensification has been shown to (Aizen & Harder 2009) and what this could mean for future reduce appropriate nesting resources for some pollinators cropping systems. Managed pollinators, mainly Apis mellifera (Kremen et al. 2002). However, some pollinating taxa can (honey bee), are heavily relied on to supplement the natural cope with moderate habitat loss and some agricultural systems pollination of agricultural crops (Potts et al. 2010). However, can provide nesting sites (Tylianakis et al. 2006; Potts et al. A. mellifera populations have been threatened by a parasitic 2010). In agriculturally dominated landscapes, patches of mite (Varroa destructor) that has spread throughout nearly all non-crop vegetation such as hedgerows may be important for major beekeeping countries (Iwasaki et al. 2015). Declines sustaining pollinator populations. Hedgerows planted with in A. mellifera are likely driven by pests, pathogens and insect-pollinated plants have been found to provide resources management choices (Smith et al. 2013) and the current state for some pollinating species (Morandin & Kremen, 2013). of wild pollinators is gaining importance (Potts et al. 2010). Yet, in New Zealand, hedgerows are typically replaced by The importance of wild pollinators for pollinating shelterbelts of trees planted predominately with Pinus radiata agricultural cropping systems has become increasingly and Cupressus macrocarpa (wind-pollinated exotic species). documented. Garibaldi et al. (2013) found that fruit set Davidson and Howlett (2010) found that only two pollinating increased twice as strongly with visitation from wild pollinators taxa (A. mellifera and Bombus spp.) were associated with than by A. mellifera and increased significantly in all of the these tree species. insect-pollinated cropping systems surveyed. Further evidence There is increasing interest in the state of wild pollinator shows that high native bee diversity is correlated with better communities both in urban parks (Webber et al. 2012) and pollination services and fruit set (Kremen et al. 2002; Klein in agro-ecosystems in New Zealand. Two New Zealand et al. 2003) and non-bee pollinators are able to provide levels studies investigated A. mellifera and wild pollinator richness, of pollination services similar to those provided by bees (Rader and compared efficiency and effectiveness of pollinators in et al. 2016). Ghazoul (2005) suggested that regional and local agricultural systems. Rader et al. (2014) found that pollinator declines of pollinators are likely affecting crop pollination. richness was lower in high intensity systems, and Rader et al. New Zealand Journal of Ecology (2018) 42(2): 262-268 © New Zealand Ecological Society.
Macdonald et al.: Landscape effects on wild pollinators 263 (2009) found that some common wild pollinator species were beside shelter belts at least 5 m in height and 30 m in length, as efficient as A. mellifera, suggesting that promoting wild that were planted with pine or macrocarpa. On each property, pollinator communities might result in services similar to measurements were taken at two replicates of each of the five those provided by A. mellifera. Although exotic pollinator landscape features, resulting in a total of four replicates for species dominated highly modified agro-ecosystems, they each feature and 20 sites across both properties. There was a cannot replace the functional composition of native pollinators, minimum distance of 50 m between sites. which could have serious implications for pollination services Two sampling methods were used at all 20 sites. The (Stavert et al. 2017). While differences in pollinator diversity flower method used potted plants of Veronica catarractae and efficiency across farming systems are informative, (syn. Parahebe catarractae) that were approximately 20 cm management to promote wild pollinator populations is likely in height with at least 40 open white flowers per plant. This to be at a local scale and involve only minor modifications species was used both because it was producing high numbers to the landscape. of flowers and because the flowers were easily accessible by Our first research aim was to investigate how local pollinators. Four potted Veronica plants were placed at each site landscape features affected wild pollinator abundance and and then observed for 10 min periods to record insect visitation species diversity on farming properties in the Canterbury to flowers. Six to eight recording periods were carried out for Region, New Zealand. Four local landscape features reflecting each site when the temperature was between 15˚C and 28˚C with contrasting histories of cultivation, species composition and wind speeds no greater than 11 km h−1. Sampling at each site vegetation structure were chosen and compared to control sites occurred at different times of the day over multiple days and (open pasture, which is the dominant landscape element in there was a minimum of 15 minutes between sampling periods at Canterbury). This research could provide information on how the same site to reduce the chance of the same pollinators being to manage local landscapes to enhance pollinator communities counted twice. During observation periods, insects that landed in order to sustain pollination services in the face of managed on the Veronica flowers were noted as potential pollinators pollinator declines and increased pollination demands, and counted. These insects were identified visually, in some particularly in land uses that negatively influence pollinator cases to species (e.g. Melanostoma fasciatum), sometimes to populations. Our second aim was to compare two different genus (e.g. Lasioglossum species) and sometimes to higher sampling methods used to measure pollinator abundances taxa (e.g. Tachinidae species). All taxa represent one species and richness (sticky traps and flower visitation). Sticky traps with the exception of Lasioglossum (two species), Bombus, have previously been used to measure pollinator community Pollenia and Tachinid which may contain a number of species. composition by Plant and Food Research Ltd in Canterbury. This may create some bias when analysing species richness. We also recorded insect visitation to flowers on potted plants However, we do not believe that this would be significant, placed at the field sites. The two methods were compared to and were unable to identify all taxa to species level without determine if choice of sampling method affects the results of trapping all pollinators. Where possible, unknown insects pollinator studies. were identified by taking photographs, to later compare to an unpublished collection of photos of local pollinators prepared by Plant and Food Research Ltd. Methods For the sticky trap method an insect trap was erected at each site and left in the field for 15 days. These traps were Pollinator measurements were carried out in February 2015 made using 4 L paint tins (80 mm diameter and 82 mm on two farming properties in Canterbury. By completing all height) wrapped in yellow plastic with a yellow base. The sampling during a single month, we avoided confounding cylinder was wrapped in clear plastic (approximately 90 mm effects of seasonal changes in pollinating taxa. The first × 570 mm), covered in “Tangle-Trap” to create a coating to property (43°54.8ʹ S, 172°6.2ʹ E) was located approximately which small insects would stick. These traps were set up on 18 km southeast of the Rakaia Township near Dorie (study stakes approximately 1 m off the ground. At the end of the 15 site approximately 308 ha), and the second property (43°33.7ʹ days the plastic strips were taken into the laboratory and the S, 171°40.1′ E) was located approximately 35 km northwest pollinator specimens were identified, sometimes to species of the Rakaia Township near the Rakaia Gorge (study site level, sometimes to genus level and sometimes to a higher approximately 210 ha). The properties were mainly used taxonomic level using an unpublished entomology pollinator for a range of cropping systems, with some pasture used for guide from Plant and Food Research Ltd. Insects smaller than grazing of stock. Delia platura (< 4 mm in length) were not counted, nor were Five local landscape features were studied. (1) Control non-pollinator species (as advised by Plant & Food Research sites: typically flat areas located in permanent pasture at least Ltd staff). 50 m from any of the other landscape features. Pasture was used as the control because it was the dominant habitat between Statistical Analysis local landscape features. (2) Bare ground: areas at least 2 m2 All statistical testing used the statistical programme R, version in size, where the ground was bare within a paddock or at a 3.2.1 (R Development Core Team 2015) and R packages lme4 fence line with minimal grass. Bare ground was used as a and vegan versions 1.1-10 and 2.4-1, respectively (Bates et feature because some native solitary bees use bare ground al. 2015; Oksansen et al. 2016). for nesting sites (Donovan 2007). (3) Biodiversity planting: Each dataset (flower and sticky trap) was analysed using located in plantings undertaken by staff of Plant & Food Generalised Linear Mixed-effects Models (GLMMs) to test Research Ltd during 2013, using a variety of native species whether landscape feature (a factor with 5 levels) significantly to promote pollinator abundances. (4) Garden: located in the affected the abundance of each pollinating taxon. To test this, gardens surrounding the main farmhouses. There were two of the glmer function was run with “property” as the random these farmhouses on each property, and in each garden there effect and “landscape feature” as the fixed effect. Poisson error was a range of native and exotic plants. (5) Shelterbelt: located distributions and their canonical log link function were used.
264 New Zealand Journal of Ecology, Vol. 42, No. 2, 2018 All taxa with an abundance of 20 or higher were run through from the dataset. These analyses were also non-significant this analysis separately. Additionally, three new variables were and are therefore not reported in more detail. created for each dataset, one for total native insect abundance, one for total insect abundance and one for number of taxa (richness), again run using the glmer function with Poisson Results distributions. To control for varying numbers of observation periods in the flower method while retaining Poisson (count) Using the flower method, 17 taxa and 928 individual pollinators data, the response variable was the total number of each taxon were recorded, which was an average of 900 insect visits per seen across all intervals at a site, then we corrected for variable hour across all taxa (Table 1). Eight of the taxa had counts numbers of 10-minute observation intervals per site using an of 20 or above and were analysed separately. The sticky trap “offset” argument in the glmer function. To remove the effect of method caught 16 taxa across 791 individual pollinators (Table differences in observation time at different landscape features, 2). Twenty or more individuals were recorded for seven taxa, in Table 1 we present mean visitors per hour. including five taxa that were common in the flower dataset. The effects of individual landscape features (levels of For both sampling methods, landscape feature was a poor this factor) were tested using the model coefficients. To predictor of pollinator richness (number of taxa). The effects investigate whether landscape feature as a whole provided on pollinator richness were small (Tables 1, 2) and rarely significant explanatory power, the analyses were re-run without significant (Appendix S1, S2 in Supplementary Materials). landscape feature. The two models were then compared using However, landscape feature was a significant factor for total a likelihood ratio test (the ANOVA function in R). When the number of insects and total number of native insects. There likelihood ratio test was non-significant we chose the simplest were significantly fewer native insects caught at each landscape model as the best fit. feature compared to the control feature in sticky traps. However, A large number of GLMM tests were run, particularly on more native insects were observed at the biodiversity planting taxon level data, which can result in an increased risk of Type and garden sites for the flower dataset. For total pollinator I error. To determine the likely influence of this problem on abundance, fewer were found at shelterbelt stations for both the results, a Bernoulli process was used to calculate the exact methods, the other three landscape features had contrasting probability of getting the number of significant results found results depending on sampling method (Table 1 & 2). in this study, given the number of trials, following Moran At the taxon level, all significant differences were in the (2003). The data were further analysed using ordinations to test direction of fewer pollinators at shelterbelt sites compared with whether there was any overall change in pollinator community control sites for both sampling methods (Table 1 & 2). The composition with regard to sampling method and landscape results for the bare-ground feature were similar, with fewer feature. This community analysis may reveal overall changes pollinators found, except that morning-evening fly (Delia in the whole community, in addition to the “single taxon at a platura) increased (Table 1 & 2). The results for biodiversity time” approach used in the GLMMs. planting and garden sites were distinctive depending on The metaMDS function from the vegan package was sampling method. There tended to be greater abundances chosen to perform non-metric multidimensional scaling per taxon using the flower method (Table 1) than with the (NMDS). This function was run in R using the Hellinger sticky trap method (Table 2). This result suggests that the two distance metric because it is recommended for species sampling methods are affected differently by nearby floral abundance data (Rao, 1995; Legendre & Gallagher 2001). resources (see Discussion); Appendix S1 gives full details of Three ordinations were run using this method. The first was the GLMM analyses. run on the entire data set to compare the sampling methods According to the Bernoulli process used to address the and then a separate ordination was run on each of the flower risk of type I error, the probability of finding this number of and sticky trap datasets to compare landscape features under significant effects (8 of 11 for the flower dataset and 8 of 10 constant sampling methods. for the sticky trap dataset) by chance was lower than 0.001 for Additionally, permutation multivariate analyses of variance both datasets, indicating that these are likely to be real effects. (PERMANOVAs) were run on all three of the ordinations using The ordination for the entire data set (stress = 0.17) the adonis function in the vegan package within R. The entire separated out the sampling methods in ordination space and dataset PERMANOVA compared the difference in community the PERMANOVA showed there was significantly different composition as a consequence of sampling method, and the community compositions measured by the two methods flower and sticky trap PERMANOVAs compared the difference (F1, 38 = 9.50, R2 = 0.20, P < 0.001) (Fig. 1a). The ordination in community composition as a consequence of landscape run on the flower dataset (stress = 0.17) showed substantial feature. This method measures distances of each site to their overlap between different landscape features in ordination centroid in multivariate community composition space, then space (Fig. 1b), but the PERMANOVA showed that there uses a permutation procedure to determine statistically if were significant systematic changes in pollinator community there was a difference in community composition. We used composition among the different landscape features (F4, 15 = the Bray-Curtis distance metric, which includes an abundance 1.59, R2 = 0.30, P = 0.002). The ordination run on the sticky trap weighting. Thus, it takes into consideration which taxa are dataset had even less distinction between landscape features. present and how abundant they are. The PERMANOVA run on the sticky trap dataset (not shown) Because the PERMANOVA for the sticky trap data was was non-significant, indicating that there was no difference non-significant, further analyses were run to explore whether between pollinator communities in different landscape features the presence of rare species was skewing the results. First, we when considering all taxa at once. re-ran the analysis using the Jaccard distance metric, which tests for only presence or absence (does not provide weighting for abundance) and secondly we re-ran the analysis after removing taxa with very low abundance (< 3 individuals)
Macdonald et al.: Landscape effects on wild pollinators 265 Table 1. The number of insects per hour visiting potted flowers at each landscape feature for each taxon. Taxa highlighted in grey had 20 or more individuals and were used in further analysis; within those rows, features in grey were significantly different from Controls (in direction shown by symbols). For statistical tests see Appendix S1. *Exotic species __________________________________________________________________________________________________________________________________________________________________ Taxon Control Bare Bio- Garden Shelter Total ground diversity belt __________________________________________________________________________________________________________________________________________________________________ Orange hoverfly (Melanostoma faciatum) 51.0 ▼27.5 56.0 ▲78.9 ▼14.0 227.4 *Striped thorax fly (Oxysarcodexia varia) 37.0 ▲60.2 ▲59.0 ▼20.1 ▼6.0 182.4 Lasioglossum species 15.0 14.0 ▲100.0 ▲46.7 0.0 175.7 *Morning evening fly (Delia platura) 9.0 ▲49.3 9.0 3.4 15.0 85.7 *Honeybee (Apis mellifera) 2.0 3.8 ▲39.0 ▲11.3 0.0 56.0 Black hoverfly (Melangyna novae-zealandiae) 10.0 4.0 13.0 9.1 7.0 43.1 Drone fly (Eristalis tenax) 5.0 3.8 ▲15.0 5.1 1.0 29.9 *Bronze thorax fly (Pollenia species) 3.0 2.0 9.0 6.1 1.0 21.1 Green soldier fly (Odontomyia species) 1.0 0.0 15.0 1.0 1.0 18.0 *Tachinid species 0.0 5.0 9.0 2.0 0.0 16.0 *Brown blowfly (Calliphora stygia) 6.0 2.8 0.0 2.0 0.0 10.6 *Euro-green blowfly (Lucilia sericata) 2.0 1.0 7.0 0.9 0.0 10.9 *Eumerus funeralis 0.0 0.0 7.0 0.0 0.0 7.0 *Euro-blue blowfly (Calliphora vicina) 0.0 0.0 1.0 2.6 1.0 4.6 Blue hoverfly (Helophilus hochstetteri) 2.0 0.0 3.0 0.0 0.0 5.0 *Bumblebee (Bombus terrestris) 0.0 1.8 1.0 0.0 1.0 3.6 *Other Bombus species 0.0 0.0 1.0 1.7 0.0 2.7 __________________________________________________________________________________________________________________________________________________________________ Native pollinator abundance 84.0 ▼ 49.3 ▲ 202.0 ▲ 140.8 ▼ 23.0 499.1 Total pollinator abundance 143.0 175.0 ▲ 344.0 ▲ 190.9 ▼ 47.0 899.7 Pollinator diversity 12 12 16 14 9 17 __________________________________________________________________________________________________________________________________________________________________ Table 2. The total number of insects caught in sticky traps at each landscape feature for each taxon. Taxa are in the same order as in Table 1; grey highlighting and symbols as for Table 1, with statistical tests in Appendix S2. *Exotic species __________________________________________________________________________________________________________________________________________________________________ Taxon Control Bare Bio- Garden Shelter Total ground diversity belt __________________________________________________________________________________________________________________________________________________________________ Orange hoverfly (Melanostoma faciatum) 51 ▼3 ▼8 ▼ 29 ▼3 94 *Striped thorax fly (Oxysarcodexia varia) 3 5 0 0 2 10 Lasioglossum species 51 ▼ 11 ▼8 ▼ 29 0 99 *Morning evening fly (Delia platura) 75 ▲ 116 61 ▲ 104 ▼ 34 390 *Honeybee (Apis mellifera) 1 2 0 2 0 5 Black Hoverfly (Melangyna novae-zealandiae) 8 2 3 5 2 20 Drone fly (Eristalis tenax) 0 1 0 0 0 1 *Bronze thorax fly (Pollenia species) 9 8 4 ▲ 20 2 43 Green soldier fly (Odontomyia species) 5 0 4 0 2 11 *Tachinid species 16 10 9 24 16 75 *Euro-green blowfly (Lucilia sericata) 1 0 0 0 0 1 *Eumerus funeralis 1 0 0 1 0 2 *Euro-blue blowfly (Calliphora vicina) 0 0 0 1 0 1 Blue hoverfly (Helophilus hochstetteri) 4 0 0 4 0 8 *Three-spotted fly (Anthomyiia punctipennis) 22 ▼3 ▼3 ▼2 0 30 Hylaeus species 0 0 0 1 0 1 __________________________________________________________________________________________________________________________________________________________________ Native pollinator abundance 119 ▼ 17 ▼ 23 ▼ 68 ▼7 234 Total pollinator abundance 247 ▼161 ▼100 222 ▼61 791 Pollinator diversity 13 10 8 12 7 16 __________________________________________________________________________________________________________________________________________________________________
266 New Zealand Journal of Ecology, Vol. 42, No. 2, 2018 (a) Flower (b) Bareground 1.0 Sticky Trap Biodiversity Control 0.5 Garden Shelterbelt 0.5 0.0 NMDS2 NMDS2 0.0 -0.5 -0.5 -1.0 -1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0 NMDS1 NMDS1 Figure 1. A non-metric multidimensional scaling ordination, using the Hellinger distance metric on (a) the entire dataset by sampling method and on (b) potted flower visitation by landscape feature. Discussion could be abundant nesting resources in an agricultural matrix for ground-nesting pollinators, suggesting that the bare ground Landscape features significantly altered both the measured sites are not essential as a nesting resource. abundance of pollinators and pollinator community The two methods were apparently measuring different composition, with different taxa responding to varying degrees; aspects of the pollinator communities, with the two datasets the results also varied according to the sampling method. showing disparate community composition in the ordination, For landscape feature effects, in the flower dataset and sometimes opposite effects in the taxon-level analyses. biodiversity plantings and gardens had higher richness and In particular, the sampling methods gave opposite results for greater pollinator abundance compared to the other landscape the garden and biodiversity planting features. One possible features. Adjacent areas with high pollinator abundance and explanation could be that the sticky traps are visited (landed diversity can lead to greater crop pollination services and fruit on) less often when there are real floral resources in close set (Garibaldi et al. 2011; Caralheiro et al. 2012; Blaauw & proximity. This difference in visitation could be due to colour Isaacs 2014). Increasing the heterogeneity and abundance not being the sole attractant for pollinators and other factors of flowering plants, particularly native species, will likely influencing pollinator behaviour (Gumbert 2000; Campbell increase the wild pollinator abundance and diversity. In turn, et al. 2010). Perhaps sticky traps are able to attract certain this could increase pollination services on Canterbury farms. pollinators from distances (being large and bright yellow), The low numbers of pollinators at shelterbelts indicate but when pollinators get closer they are more interested in that pine and macrocarpa shelterbelts (both of which are real floral resources nearby with more ‘attractive’ attributes. wind-pollinated) provide poor habitat for insect pollinators, Likewise, in the gardens there was a diverse range of floral and are potentially reducing the local abundance of wild resources (varying in colours, shapes, and sizes), which could pollinators on Canterbury farms. Improving habitat for insects also have reduced the appeal of both, the flower and trap in hedgerows can be an important conservation tool for methods, for some pollinator species. For example, Bombus pollinator enhancement (Morandin & Kremen, 2013). Thus, species were often observed visiting Lavandula angustifolia shelterbelts could be planted with floral understory targeted in one of the homestead gardens, but only low numbers were towards wild pollinators, and new hedgerows could be planted observed visiting the experimental potted Veronica flowers predominately with native insect-pollinated species. However, and none were caught in the sticky traps. We consider the hedgerows can act as a barrier for pollinators between adjacent flower method to be a better measure of local pollinator areas (Wratten et al. 2003: Klaus et al. 2015), which should community composition for this research. More pollinators be considered in the conservation of pollinator communities were recorded using the flower method with most taxa found and crop pollination. at higher abundances. The flower results also matched casual Lasioglossum was the only genus recorded in this study that observations at the sites showing more pollinator activity uses bare ground as a nesting resource. There was no significant around biodiversity plantings and gardens (not less, as the difference between Lasioglossum abundance at bare ground sticky traps found). sites compared with control sites using the flower method. The flower method measured the actual flower visitors Donovan (2007, p. 134) notes that Lasioglossum sordidum to small, white native flowers. Thus, this sampling method nests are usually reported as being in bare ground. However, a provided targeted monitoring for pollinators of a specific study in the Christchurch Botanic Gardens found them nesting native plant. The flower method required approximately 60 in lawns among grass (Bennet et al. 2018). Therefore, there hours of sampling time to gather the field data. The sticky
Macdonald et al.: Landscape effects on wild pollinators 267 trap method catches insects in the vicinity of the trap because Carvalheiro LG, Seymour CL, Nicolson SW, Veldtman R either: (1) the insects are blown into the traps by the wind or 2012. Creating patches of native flowers facilitates crop (2) the insects choose to land on the trap, perhaps attracted to pollination in large agricultural fields: mango as a case the large yellow shape. The glue used to trap insects is less study. Journal of Applied Ecology 49: 1373–1383. effective on large insects (≥ honeybees); thus, sticky traps are Collins KE, Doscher C, Rennie HG, Ross JG 2013. The biased towards smaller insects. The sticky trap method requires effectiveness of riparian ‘restoration’ on water quality-A less time in the field (approximately 25 hours to place traps case study of lowland streams in canterbury, New Zealand. out and retrieve them), but does require more time in the lab Restoration Ecology 21: 40–48. to separate, identify and count insects. The flower method Davidson MM, Howlett BG 2010. Hedging our bets: choosing could be easily adapted to attract a wider range of pollinators hedgerow plants to enhance beneficial insects to optimise by creating a portable planter box with species that vary in crop pollination and pest management on Canterbury flower colour, size and structure. farms. Auckland, The New Zealand Institute for Plant The flower method provided evidence that biodiversity and Food Research Limited. 25 p. plantings attract greater numbers of pollinators. These results De Groot RS, Wilson MA, Boumans RMJ 2002. A typology for indicate that targeted native planting is a useful tool to enhance the classification, description and valuation of ecosystem wild pollinator communities on farms in Canterbury. If native functions, goods and services. Ecological Economics 41: plantings are used in farm management, future research could 393–408. investigate three questions. (1) Can native plantings provide Donovan BJ 2007. Apoidea (Insecta: Hymenoptera). Fauna of an increase in crop yields in adjacent fields? (2) How close New Zealand 57. Lincoln, Manaaki Whenua Press. 295 p. do these plantings need to be to enhance pollinators and crop Garibaldi LA, Steffan-Dewenter I, Kremen C, Morales JM, yields? (3) Do the size and shape of these plantings have an Bommarco R, Cunningham SA, Carvalheiro LG, Chacoff effect on pollinators and crop yields? Carvalheiro et al. (2012) NP, Dudenhöffer JH, Greenleaf SS, Holzschuh A, Isaacs found that areas of native plantings as small as 25 m2 could R, Krewenka K, Mandelik Y, Mayfield MM, Morandin enhance pollinator-dependent crop production if combined LA, Potts SG, Ricketts TH, Szentgyörgyi H, Viana BF, with other land management. Biodiversity plantings are a Westphal C, Winfree R, Klein AM et al. 2011. Stability tool that is already used on farmland to support ecosystem of pollination services decreases with isolation from services. For example, native riparian plantings are used to natural areas despite honey bee visits. Ecology Letters improve water quality (Collins et al. 2013) and other native 14: 1062–1072. plantings have been used to support natural biological control Garibaldi LA, Steffan-Dewenter I, Winfree R, Aizen MA, (Walton & Isaacs 2011). The results from this study further Bommarco R, Cunningham SA, Kremen C, Carvalheiro support biodiversity plantings as a tool on farmlands to enhance LG, Harder LD, Afik O, Bartomeus I, Benjamin F, Boreux species providing pollination services. V, Cariveau D, Chacoff NP, Dudenhöffer JH, Freitas BM, Ghazoul J, Greenleaf S, Hipólito J, Holzschuh A, Howlett B, Isaacs R, Javorek SK, Kennedy CM, Krewenka Acknowledgements KM, Krishnan S, Mandelik Y, Mayfield MM, Motzke I, Munyuli T, Nault BA, Otieno M, Petersen J, Pisanty G, We thank Brad Howlett and Samantha Read for help with Potts SG, Rader R, Ricketts TH, Rundlöf M, Seymour logistics and insect identification, and thank MBIE C11X1309 CL, Schüepp C, Szentgyörgyi H, Taki H, Tscharntke T, Bee Minus to Bee Plus and Beyond: Higher Yields From Vergara CH, Viana BF, Wanger TC, Westphal C, Williams Smarter, Growth-focused Pollination Systems, MPI Sustainable N, Klein AM et al. 2013. Wild pollinators enhance fruit Farming Fund project (12–015, Building Better Biodiversity set of crops regardless of honey bee abundance. Science on Cropping Farms) and The New Zealand Institute for Plant 340: 1608–1611. & Food Research Limited for funding. Garibaldi LA, Carvalheiro LG, Leonhardt SD, Aizen MA, Blaauw BR, Isaacs R, Kuhlmann M, Kleijn D, Klein AM, Kremen C, Morandin L, Scheper J, Winfree R 2014. From References research to action: enhancing crop yield through wild pollinators. Frontiers in Ecology and the Environment Aizen MA, Harder LD 2009. The global stock of domesticated 12: 439–447. honey bees is growing slower than agricultural demand Ghazoul J 2005. Buzziness as usual? Questioning the global for pollination. Current Biology 19: 915–918. pollination crisis. Trends in Ecology & Evolution 20: Bates D, Maechler M, Bolker B, Walker S 2015. Fitting Linear 367–373. Mixed-Effects Models Using lme4. Journal of Statistical Gumbert A 2000. Color choices by bumble bees (Bombus Software, 67: 1–48. terrestris): innate preferences and generalization after Bennet DG, Kelly D, Clemens J 2018. Food plants and foraging learning. Behavioral Ecology and Sociobiology 48: 36–43. distances for the native bee Lasioglossum sordidum in Iwasaki JM, Barratt BIP, Lord JM, Mercer AR, Dickinson KJM Christchurch Botanic Gardens. New Zealand Journal of 2015. The New Zealand experience of varroa invasion Ecology 42: 40–47. highlights research opportunities for Australia. Ambio Blaauw BR, Isaacs R 2014. Flower plantings increase wild 44: 694–704. bee abundance and the pollination services provided to a Klaus F, Bass J, Marholt L, Müller B, Klatt B, Kormann pollination-dependent crop. Journal of Applied Ecology U 2015. Hedgerows have a barrier effect and channel 51: 890–898. pollinator movement in the agricultural landscape. Journal Campbell DR, Bischoff M, Lord JM, Robertson AW 2010. of Landscape Ecology 8: 22–31. Flower color influences insect visitation in alpine Klein AM, Steffan–Dewenter I, Tscharntke T 2003. Fruit set of New Zealand. Ecology 91: 2638–2649. highland coffee increases with the diversity of pollinating
268 New Zealand Journal of Ecology, Vol. 42, No. 2, 2018 bees. Proceedings of the Royal Society of London B: pollination services: Are there general patterns? Ecology Biological Sciences 270: 955–961. Letters 11: 499–515. Klein AM, Vaissière BE, Cane JH, Steffan-Dewenter I, Smith KM, Loh EH, Rostal MK, Zambrana-Torrelio CM, Cunningham SA, Kremen C, Tscharntke T 2007. Mendiola L, Daszak P 2013. Pathogens, pests, and Importance of pollinators in changing landscapes for world economics: Drivers of honey bee colony declines and crops. Proceedings of the Royal Society B: Biological losses. EcoHealth 10: 434–445. Sciences 274: 303–313. Stavert JR, Pattemore DE, Gaskett AC, Beggs JR, Bartomeus I Kremen C, Williams NM, Thorp RW 2002. Crop pollination 2017. Exotic species enhance response diversity to land-use from native bees at risk from agricultural intensification. change but modify functional composition. Proceedings of Proceedings of the National Academy of Sciences of the the Royal Society B: Biological Sciences 284: 20170788. United States of America 99: 16812–16816. Tylianakis JM, Klein AM, Lozada T, Tscharntke T 2006 Spatial Legendre P, Gallagher E 2001. Ecologically meaningful scale of observation affects α, β and γ diversity of cavity- transformations for ordination of species data. Oecologia nesting bees and wasps across a tropical land-use gradient. 129: 271–280. Journal of Biogeography 33: 1295–1304. Moran MD 2003. Arguments for rejecting the sequential Webber CJ, Peterson AJ, Kelly D, Clemens J 2012. Native and Bonferroni in ecological studies. Oikos 100: 403–405. exotic flower visitors in the Christchurch Botanic Gardens Morandin LA, Kremen C 2013. Hedgerow restoration promotes and their contrasting plant preferences. New Zealand pollinator populations and exports native bees to adjacent Natural Sciences 37: 36–48. fields. Ecological Applications 23: 829–839. Walton NJ, Isaacs R 2011. Influence of native flowering plant Oksanen J, Blanchet, FG, Friendly M, Kindt R, Legendre strips on natural enemies and herbivores in adjacent P, McGlinn D, Minchin PR, O’Hara RB, Simpson GL, blueberry fields. Environmental Entomology. 40: 697–705. Solymos P, Stevens MHH, Szoecs E, Wagner H 2016. Wratten SD, Bowie MH, Hickman JM, Evans AM, Sedcole vegan: Community Ecology Package. R package version JR, Tylianakis JM 2003 Field boundaries as barriers to 2.4-1. movement of hover flies (Diptera: Syrphidae) in cultivated Potts SG, Biesmeijer JC, Kremen C, Neumann P, Schweiger land. Oecologia 134: 605–611. O, Kunin WE 2010. Global pollinator declines: Trends, impacts and drivers. Trends in Ecology and Evolution Received 13 November 2017; accepted 13 April 2018 25: 345–353. Editorial board member: David Pattemore and George Perry R Development Core Team 2015. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rader R, Howlett BG, Cunningham SA, Westcott DA, Supplementary Material Newstrom-Lloyd LE, Walker MK, Teulon DA, Edwards W 2009. Alternative pollinator taxa are equally efficient Additional supporting information may be found in the online but not as effective as the honeybee in a mass flowering version of this article: crop. Journal of Applied Ecology 46: 1080–1087. Rader R, Bartomeus I, Tylianakis JM, Laliberté E 2014. Appendix S1. Summary of Generalised Linear Mixed Model The winners and losers of land use intensification: (GLMM) results, using the data from potted flower visitation, Pollinator community disassembly is non-random and testing abundance of each taxon against landscape feature, alters functional diversity. Diversity and Distributions with property as the random effect. 20: 908–917. Appendix S2. Summary of Generalised Linear Mixed Model Rader R, Bartomeus I, Garibaldi LA, Garratt MP, Howlett BG, (GLMM) results, using the data from pollinators caught in Winfree R, Cunningham SA, Mayfield MM, Arthur AD, sticky traps, testing abundance of each taxon against landscape Andersson GKS, , Bommarco R, Brittain C, Carvalheiro feature, with property as the random effect. LG, Chacoff NP, Entling MH, Foully B, Freitas BM, Gemmill-Herren B, Ghazoul J, Griffin SR, Gross CL, The New Zealand Journal of Ecology provides online Herbertsson L, Herzog F, Hipólito J, Jaggar S, Jauker F, supporting information supplied by the authors where this Klein A, Kleijn D, Krishnan S, Lemos CQ, Lindström may assist readers. Such materials are peer-reviewed and SAM, Mandelik Y, Monteiro VM, Nelson W, Nilsson copy-edited but any issues relating to this information (other L, Pattemore DE, de O. Pereira N, Pisanty G, Potts SG, than missing files) should be addressed to the authors. Reemer M, Rundlöf M, Sheffield CS, Scheper J, Schüepp C, Smith HG, Stanley DA, Stout JC, Szentgyörgyi H, Taki H, Vergara CH, Viana BF,Woyciechowski Met al. 2016. Non-bee insects are important contributors to global crop pollination. Proceedings of the National Academy of Sciences of the United States of America 113: 146–51. Rao CR 1995. A review of canonical coordinates and an alternative to correspondence analysis using Hellinger distance. Questiió: Quaderns d’Estadística, Sistemes, Informatica i Investigació Operativa 19: 23–63. Ricketts TH, Regetz J, Steffan-Dewenter I, Cunningham SA, Kremen C, Bogdanski A, Gemmill-Herren B, Greenleaf SS, Klein AM, Mayfield MM, Morandin LA, Ochieng’ A, Potts SG, Viana BF 2008. Landscape effects on crop
You can also read